Hi, first necessary thing is using "cartoon" prompt as I explained. The rest of the prompt is a free ride though - the only SFW example I posted is with a basic prompt "cartoon, Emma Watson". The rest is generated with combinations of tags that the model learned from autotagged captions (WD14 style), you should be able to find them in any LoRA manager, but for example things like "cartoon, 1boy, 1girl, sex" worked for me during the test, also full sentence version like "cartoon, a girl with hand in panties sitting on bed in front of a boy" worked too. If you can post your prompts, I can test that at my side too.

Oh right, my bad for forgetting to switch the checkpoint version on the model page. Thanks for spotting that :)

@aa9739614 Thanks for the feedback :)

@Santaonholidays Faces in distance generally require use of some version of ADetailer which these samples do not have (can't run that on my poor GPU). Genitals are kind of a hit and miss especially if you put both sexes in the same image without using something like Regional Prompter. For the sample images, I deliberately chose poses that are pretty hard to generate without these tools, as opposite to just close-up portraits of faces looking into the camera (like the one Emma Watson sample). I would suggest just downloading and testing the model yourself with your own settings and prompts, my samples are not cherry picked and only show what the model is supposed to do.